I had it a few times already (though not that much in the many years I use this already), that some entries were not ended on a second device, because the end message was received before the start message.
I added logic to compensate for this, but it is not good enough because I just had the issue again -> so we need to write some unit tests for the state handling to test if the logic is resistant to out of order delivery of states. - we could also explore if it would make sense to add some p2p state management library like y.js
I had it a few times already (though not that much in the many years I use this already), that some entries were not ended on a second device, because the end message was received before the start message.
I added logic to compensate for this, but it is not good enough because I just had the issue again -> so we need to write some unit tests for the state handling to test if the logic is resistant to out of order delivery of states. - we could also explore if it would make sense to add some p2p state management library like y.js